Senior AI Scientist

The AI Accelerator is a brand-new London‑based hub within Computational Innovation (CI), a global organisation comprising computational biology, human genetics, data excellence and AI expertise. The purpose of CI’s AI Accelerator is to provision production‑quality, versatile, foundational biomedical AI capabilities that can be adapted and deployed to improve and accelerate portfolio decision‑making and increase the probability of success, by furthering understanding of the biology driving patient outcomes and identifying mechanisms involved in disease.

A core component of the AI department is AI Systems, a team focused on designing, building and deploying versatile biomedical foundation models that, through adaptation, can enhance human understanding of disease biology and help identify potential targets, biomarkers and patient segments for further research. AI Systems will exploit neural‑based methods to integrate data and impute and infer across the biomedical landscape, including electronic health records and medical imaging to support patient segmentation, omics data to understand gene regulation and identify novel therapeutic concepts, and prediction of transcriptional changes for a given disease‑causing variant.

We are seeking a Senior AI Scientist to join the Accelerator’s AI Systems team and deliver next‑generation, foundational AI capabilities to support discovery and development of innovative medicines. You will be an experienced, independent AI Scientist within AI Systems, owning specific research workstreams within the broader research direction set by the Senior Staff AI Scientist. You will design and run experiments to test model architectures and training approaches, interpret results rigorously and iterate toward validated prototypes. You will work alongside ML Engineers who contribute production engineering perspectives to early research discussion, ensuring architectural choices are both scientifically sound and practically implementable. You will also work closely with stakeholders to understand the downstream use cases your research should serve, and with the Senior Staff AI Scientist who provides mentorship and research guidance. You are expected to operate with increasing independence, developing your own research instincts and growing your contribution to architectural and technical direction over time.

Key Responsibilities

• Develop and validate novel approaches for learning from multi‑omics, clinical and imaging data, producing reproducible prototypes and documentation for those that will go into production.
• Contribute ideas to architectural discussions and research direction, growing toward greater ownership of research decisions over time.
• Work in close partnership with ML Engineers throughout the research and implementation process.
• Collaborate with stakeholders to understand their needs and ensure models are fit for purpose.
• Stay current with the latest advances in foundation models, multi‑modal learning and biomedical AI, bringing relevant developments to the team's attention.
• Publish research findings in peer‑reviewed ML and biomedical AI venues, contributing to the team's scientific credibility and external profile.

Required Qualifications

• PhD in machine learning, computer science, computational biology, bioinformatics or a related quantitative field.
• Strong hands‑on experience with deep learning and foundation model architectures, pre‑training, transfer learning, fine‑tuning.
• Experience working with at least one biomedical data modality, e.g. multi‑omics, clinical or imaging data in an ML context.
• Strong software engineering skills, version control, testing, documentation.
• Familiarity with distributed training frameworks and large‑scale data processing.
• Familiarity with challenges specific to biomedical AI such as small sample sizes, batch effects, data integration.
• Experience collaborating with domain experts to understand requirements and validate results.

Preferred Qualifications

• Experience with multi‑modal foundation models or cross-modal transfer learning.
• Strong understanding of at least two biomedical data modalities.
• Previous work in drug discovery, healthcare or related application domains.
• Track record of publishing/open‑source contributions or widely used model releases.